Article Figures & Data
Figures
- Fig. 1.
A: photograph of a histological parasagittal section in V2 (see text for explanations). B: receptive fields (RFs) plotted for monkey P. The dotted RFs labeled V1, V2, and V3 were plotted along a single mapping penetration. The RFs labeled 1, 2, and 3 correspond to neurons recorded on 3 different electrodes positioned in V2 in another mapping penetration. The black/gray RFs labeled V1 and V2 were obtained during the inactivation experiment. See text. C: picture of the inactivation device that we used for the recordings in monkey P. The picture was taken after the experiment. A similar device was used with monkey N. A device with 5 microelectrodes was used with monkey Q. The spaces between the elements were then larger, as 200 mM GABA was used instead of 100 mM. D: plot of a frontal view of the inactivation device shown in C. The micropipettes tips (small black disks) are separated by distances inferior to 1 mm. All measures were done under a microscope equipped with a graticule before the experiment and checked after the experiment. Small gray disks represent the tips of the microelectrodes. The distances from each microelectrode to their closest micropipette(s) are indicated [calledX coordinate, the lateral distance in Hupé et al. (1999)]. The Z distance is the distance from the tip of the microelectrodes to the plane of the tips of the micropipettes. Minus means below (i.e., in the superficial layers of V2 once the device is well positioned). Microelectrode tips E1 and E3 are indicated by the arrows 1 and 2, respectively, inC. The light gray disks around the micropipettes represent the calculated inactivation zones following an injection of 50 nl of 100 mM GABA (see Hupé et al. 1999, Fig. 9) at, respectively, the Z position of about 0 or 700 μm (i.e., for a lateral distance of about 500 μm from the tip of the micropipette), and the Z position of about −400 or 1,100 μm (i.e., for a lateral distance of about 350 μm). If the device was optimally positioned in V2, the small disks would therefore represent the regions of inactivation at the very top and bottom borders of V2, and the large disks the mean region of inactivation around the pipettes. The large black disk behind the inactivation disks represents a rather homogeneously inactivated region of 2-mm diam, which is the minimal extent of V2 region that we consider to have fully inactivated. Note also that 50 nl of GABA is not a priori sufficient to inactivate the neurons recorded by the microelectrode E3, which is 675 μm above the tip of the pipettes. As during the experiment we injected GABA until these neurons were inactivated (see Fig. 2), we are confident that our inactivations were always larger than those represented here.
- Fig. 2.
Example of inactivation of V2. Multi-unit activity recordings of the V2 neurons were obtained with the 3 microelectrodes E1, E2, and E3 of the inactivation device plotted on Fig. 1 D, while the V1 neuron PBL11 was tested (see Fig. 3, bottom peristimulus time histograms (PSTHs); see Fig. 1 B for the plots of the RF of V1 and V2 neurons). The device was already for 30 h in V2 when this recording was carried out. Four successive injections of 25 nl, 100 mM GABA in each micropipette were achieved, so the total volume of 100 mM GABA injected in V2 was 600 nl. The recording started after the second injection. Note the complete disappearance of neural activity on E3, which is the furthest away from the micropipettes. Some residual activity can be observed on E2; but the evoked activity is negligible compared with what was recorded during the control, and what looks like spontaneous activity was in fact just electronic noise due to the communication between the computer and the Harvard pump connected to the micropipettes, as could be checked off-line. Such electrical artifacts could never affect the recordings of V1 neurons, as they were always rejected by the spike sorting device (MSD). The GABA recording was done 9 min after the control recording and the recovery recording 40 min after the GABA recording. Twenty repetitions of the sequence were carried out in each condition (recording time = 3 min, 30 s). Bin width = 100 ms. One SE is plotted below and above the mean response obtained during the 500 ms of stimulus presentation for each stimulus.
- Fig. 3.
Mean responses ±1 SE to the 7 stimuli depicted at thebottom of the figure for the 6 neurons for which the response to the single bar (stimulus C) significantly decreased during GABA. Left: responses during control.Right: responses during V2 inactivation (GABA). The responses to the center, the 3 center/surround, and then the 3 surround only stimuli are shown. The neurons were classified as a function of their differential response to the 4 1st stimuli (seemethods). In each box, we printed the name of the neuron, then the class they belong to for the surround orientation comparison set (comparisons of the responses to the 3 1st stimuli), and then for the spatial configuration comparison set (response to the stimuli C, C/S, and C/l).
- Fig. 4.
Time course of the effects of V2 inactivation on V1 neurons. Bin width = 20 ms. A: example of a significant decrease of response (−51%, P = 0.003) observed on a neuron with an early response (latency = 46 ms). Case pbl11, single unit. B: example of a significantly decreased response (−37%, P = 0.022), observed on a neuron with a late response (latency = 82 ms). Case pcc14, multi-unit.C: the responses of the 6 neurons for which theon response was significantly decreased by V2 inactivation were pooled. The recovery was recorded for only 5 neurons. Note that control and recovery traces are almost perfectly superimposed. There is no change of the mean peak off response. D: measure of the time course of the decrease of response due to V2 inactivation. The 1st significant bin is the 2nd one (20–40 ms after response onset), when tested with the classical (not MCP) Wilcoxon test; 4/6 neurons showed a decrease of response also in the 1st bin (0–20 ms) after response onset. Significant values (P < 0.05) could be obtained only when at least 5 neurons showed a decrease in a given bin. This histogram shows the excitatory contributions of feedback connections from V2 to theon responses of V1 neurons.
- Fig. 5.
Histogram of the latencies of the 61 V1 neurons responses to C that were recorded and tested during V2 inactivation. The latencies of the significantly affected neurons are shown. Note that several neurons affected by V2 inactivation have short latencies (both distributions are not statistically different, which can be explained by the small sample of affected neurons).
- Fig. 6.
Breakdown of the 64 neurons sample into 5 main classes and 5 mixed classes. See methods for the description of the classification criteria.
- Fig. 7.
No effect of V2 inactivation on the surround nonspecific suppression.A: normalized population responses. Responses were normalized to each neuron's response to the optimal center bar measured during the control period; these values were then averaged for 22 neurons (20 GS cells and 2 S cells, see text). B: stimuli. C: time course of the surround modulation. The responses of the 22 neurons were temporally aligned on the onset of response, with a 5-ms precision. Responses in successive bins were then added, to obtained 20-ms binwidth histograms. The amplitude was normalized to the peak response to the stimulus C measured during the control period. The traces are shown for the optimal stimulus, the center/surround stimuli, and the surround-only stimuli (bottom traces, here and in all the other plots). Note that the mean suppression is almost complete and present from the 1st bin of response. D: same as B, but during inactivation of V2 by GABA.
- Fig. 8.
No effect of V2 inactivation onto the orientation contrast enhancement. Same conventions as in Fig. 7. The responses of 12 OC cells were pooled together for all the stimuli except for the stimulus C, for which the responses of 2 neurons were statistically different during the 2 controls. These responses were rejected after the normalization had been done. The mean response to the stimulus C of only 10 cells is therefore plotted. A: normalized population responses.B: stimuli. C and D: time course histograms. E and F: histograms of differences between the responses to the stimuli C/S and C/S′. The difference of normalized response C/S′-C/S was computed for each neuron, and then computed in a population histogram, for control and then during V2 inactivation. One SE is plotted below and above the mean response, so one can directly read the level of significance of pairedt-tests done between the responses to both stimuli (assuming a Normal distribution). The bin called “0” is in fact the last bin before response onset (−20 to 0 ms). The horizontal dark bar below the histogram indicates which bins were significant (P < 0.05) when a MCP Wilcoxon exact test was done. The first significant bin is the 4th one (40–60 ms after response onset, see text). Only bins up to 200 ms after response onset were tested.
- Fig. A1.
No effect of V2 inactivation on other surround modulations. See text. Same conventions as in Fig. 7. A and B: population histograms for the neurons for which the responses to the center/surround stimuli are significantly larger than the response to the center-only stimuli. C and D: population histogram for the neurons for which the response to C/S is significantly larger than the response to C/S′.
- Fig. A2.
No effect of V2 inactivation on the neuron responses that are not modulated by the surround. See text. Same conventions as in Fig. 7.A and B: population histogram for the 20 NM neurons. C and D: population histogram for NM neurons for which the first 100 ms of response is significantly suppressed by the surround. E and F: population histogram for the neurons for which only the late part of the response (after 100 ms) increased with the surround stimuli (note: these neurons were not necessarily NM when tested over 500 ms).
StumbleUpon
